linux-kernel - Re: [PATCH] mm: compare MIGRATE

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <xa1tr4mjpo80.fsf@mina86.com>
Date:	Fri, 21 Dec 2012 13:46:23 +0100
From:	Michal Nazarewicz <mina86@...a86.com>
To:	Minchan Kim <minchan@...nel.org>
Cc:	Andrew Morton <akpm@...ux-foundation.org>, linux-mm@...ck.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] mm: compare MIGRATE_ISOLATE selectively

> On Thu, Dec 20, 2012 at 04:49:44PM +0100, Michal Nazarewicz wrote:
>> Perhaps “is_migrate_isolate” to match already existing “is_migrate_cma”?

On Fri, Dec 21 2012, Minchan Kim wrote:
> Good poking. In fact, while I made this patch, I was very tempted by renaming
> is_migrate_cma to cma_pageblock.
>
>         is_migrate_cma(mt)
>
> I don't know who start to use "mt" instead of "migratetype" but anyway, it's
> not a good idea.
>
>         is_migrate_cma(migratetype)
>
> It's very clear for me because migratetype is per pageblock, we can know the
> function works per pageblock unit.
>
>> Especially as the “mt_isolated_pageblock” sound confusing to me, it
>> implies that it works on pageblocks which it does not.
>
> -ENOPARSE.
>
> migratetype works on pageblock.

migratetype is a number, which can be assigned to a pageblock.  In some
transitional cases, the migratetype associated with a page can differ
from the migratetype associated with the pageblock the page is in.  As
such, I think it's confusing to add “pageblock” to the name of the
function which does not read migratetype from pageblock but rather
operates on the number it is provided.

> I admit mt is really dirty but I used page_alloc.c already has lots of
> mt, SIGH.

I don't really have an issue with “mt” myself, especially since the few
times “mt” is used in page_alloc.c it is a local variable which I don't
think needs a long descriptive name since context is all there.

> How about this?
>
> 1. Let's change all "mt" with "migratetype" again.
> 2. use is_migrate_isolate and is_migrate_cma for "migratetype".
> 3. use is_migrate_isolate_page instead of page_isolated_pageblock for
>    "page".

Like I've said.  Personally I don't really think 1 is needed, but 2 and
3 look good to me.

-- 
Best regards,                                         _     _
.o. | Liege of Serenely Enlightened Majesty of      o' \,=./ `o
..o | Computer Science,  Michał “mina86” Nazarewicz    (o o)
ooo +----<email/xmpp: mpn@...gle.com>--------------ooO--(_)--Ooo--

Content of type "application/pgp-signature" skipped